摘要
针对基于关系的好友推荐中偏离共同兴趣以及推荐好友数量不足的问题,将数据挖掘中FP-Growth关联规则算法应用于社交网络好友推荐中,对用户间的相互关注关系进行深度挖掘,将不同用户同时被关注的事件作为一个项集,挖掘其频繁模式,再根据设定支持度,推荐用户感兴趣Top-N组合好友。63641条实验结果表明,算法具有良好的性能,可实现较高的召回率与准确率。
In view of the problem of deviation from common interests and insufficient number of recommended friends in relationship-based friend recommendation,FP-Growth association rule algorithm in data mining is applied to social network friend recommendation to conduct deep mining of mutual concern among users.Taking events that different users are concerned about at the same time as an item set,their frequent patterns are mined.Top-N group friends whom users are interested in are then recommended according to the set support degree.Experimental results show that the algorithm has good performance and can achieve high recall and accuracy.
作者
熊才权
陈曦
XIONG Caiquan;CHENG Xi(School of Computer Science,Hubei Univ.of Tech.,430068,China)
出处
《湖北工业大学学报》
2020年第1期33-37,共5页
Journal of Hubei University of Technology
基金
国家重点研发计划项目(2017YFC1405403)
国家自然科学基金(61075059)
湖北工业大学绿色工业科技引领计划项目(CPYF2017008)。